Sweep: area

Varies: area long edge (m) — 1500 → 2500 m, in five points.

← Benchmark index · Metrics · Methodology

Provenance — measured at v1.3.0. Every number on this page was produced by the campaign runs named below, all of which predate the v1.3.0 tag (19009be). No commit between v1.2.0 and that tag changes routing behaviour, so checking out v1.3.0 reproduces these cells. The default branch will not: it has since changed the pheromone weighting (#327, betaAnts/ betaData 2.0 → 20) and the per-seed RNG stream assignment (#352), either of which moves measured values. See Provenance and #365.

Restated on CBR sources (#521). The pinned table below was measured at half the documented offered load. A 20-seed re-measure on main is under Restated on CBR sources at the end of this page.

OLSR's PDR column is inflated (#510). These cells predate the fix that counts a send stock OLSR refuses at the source as an offered packet (PR #511, f480d0ae). Before it, OLSR's PDR was computed over the ticks it had a route for. This page stays pinned to v1.3.0 and is not re-measured. The offered-based value below is estimated from this table's own columns, as thrput_olsr / thrput_aodv × PDR_aodv. On these aggregate rows the same estimate reproduces AntHocNet's and DSDV's reported PDR to within ~0.6 pp, which is the throughput rounding.

area long edge (m)	OLSR PDR as published	offered-based estimate	DSDV PDR
1500	79.3	≈ 75.2 (-4.1)	70.8
1900	73.6	≈ 67.1 (-6.5)	63.5
2100	70.1	≈ 62.0 (-8.1)	60.4
2300	68.7	≈ 59.1 (-9.6)	56.5
2500	67.0	≈ 55.8 (-11.2)	54.0

The correction reaches 11.2 pp. Where it brings OLSR within about 2 pp of DSDV, do not quote an OLSR-over-DSDV ordering from this page. AntHocNet, AODV and DSDV are unaffected.

What it varies

Reproduces Fig. 1 of the AntHocNet paper (Di Caro/Ducatelle/Gambardella, PPSN VIII 2004, §4): the base scenario's long edge is extended from 1500 m to 2500 m while node count and traffic stay fixed, so paths get longer and the network gets sparser. Unlike the paper (AODV only), every baseline (AODV/OLSR/DSDV) is run on identical realisations.

Defined as SWEEPS["area"] in ns3/tools/run-scenarios.py (scenario=paper, areaX ∈ {1500, 1900, 2100, 2300, 2500}).

How it is produced

Sweeps are too heavy for the per-merge benchmarks workflow, which runs only the discrete scenarios. They come from the manual Scenario matrix + charts workflow (scenario-matrix.yml), which renders sweep-area.png into docs/benchmarks/ and uploads the classified CSV; the table below is filled in by pointing update-benchmarks.py at that CSV.

Raw sweep data rescued from expired artifacts lives in ../campaign/. The published table below comes from the 2026-08-03 20-seed re-run (runs 30850317400 / 30850325638 / 30850333760 / 30850342363 / 30850350287, one point per job, 3.42-opt image, 900 s, range/disk PHY), rescued as ../campaign/308503*-run.csv and summarizable with the benchmark-results skill's sweep_summary.py. The superseded pre-#88/#169 5-seed data remains at 30031902395-area-disk.csv.

Results

Post-#88/#169, 20 seeds, 95% CIs. These numbers include the T_hop = 3 ms fix (#88, PR

167) and the reactive-hop-cap removal

(#169, PR #170), and meet the methodology runs floor. Versus the superseded 5-seed data, AntHocNet's PDR moved +1.1 to +13.3 pp across the sweep (baselines moved ≤ ~1 pp, consistent with the 5→20 run-count change) — the re-run existed precisely because #88/#169 invalidated the old numbers. The delay99 column carries the known #21 tail (remediation tracked in #308). Do not quote path_div_*/entropy from these cells (#230 window caveat).

Sweep area — mean of 20 run(s) per point, every baseline on identical realisations; ± is the 95% CI half-width (#293). Generated by run-scenarios.py; chart by make-charts.py.

area long edge (m)	protocol	PDR % ±95	mean delay (ms) ±95	99th delay (ms) ±95	throughput (kbps)	NRL ±95	jitter (ms)	dOff90 (ms)
1500	anthocnet	95.0 ± 0.4	57.8 ± 1.9	878.0 ± 18.8	6.24	40.325 ± 0.60	93.59	322.2
1500	aodv	84.7 ± 0.7	32.5 ± 1.6	489.8 ± 26.1	5.59	54.292 ± 1.13	49.03	inf
1500	olsr	79.3 ± 0.9	7.6 ± 0.6	25.5 ± 1.1	4.96	5.095 ± 0.10	10.60	inf
1500	dsdv	70.8 ± 1.1	15.4 ± 1.2	423.1 ± 119.9	4.70	26.486 ± 0.37	25.10	inf
1900	anthocnet	93.3 ± 0.5	66.7 ± 2.2	930.4 ± 16.5	6.12	41.114 ± 0.61	105.81	547.4
1900	aodv	81.5 ± 1.0	31.2 ± 1.8	501.8 ± 47.6	5.38	48.982 ± 1.08	44.74	inf
1900	olsr	73.6 ± 0.8	7.9 ± 0.7	28.3 ± 1.3	4.43	5.894 ± 0.09	10.50	inf
1900	dsdv	63.5 ± 1.0	15.7 ± 1.2	321.3 ± 113.7	4.22	28.476 ± 0.53	23.89	inf
2100	anthocnet	92.1 ± 0.4	72.2 ± 3.2	959.4 ± 13.6	6.05	42.481 ± 0.69	114.14	762.3
2100	aodv	80.8 ± 0.9	32.1 ± 1.3	551.6 ± 40.7	5.33	46.659 ± 0.98	45.14	inf
2100	olsr	70.1 ± 1.0	7.2 ± 0.5	28.6 ± 1.0	4.09	6.476 ± 0.14	8.97	inf
2100	dsdv	60.4 ± 1.0	14.3 ± 1.1	239.2 ± 88.8	4.01	29.355 ± 0.53	21.74	inf
2300	anthocnet	90.9 ± 0.5	75.7 ± 3.6	967.0 ± 13.7	5.96	42.992 ± 0.77	118.24	inf
2300	aodv	78.8 ± 0.9	35.5 ± 2.4	715.4 ± 79.8	5.20	43.966 ± 1.00	49.27	inf
2300	olsr	68.7 ± 1.3	7.8 ± 0.7	29.6 ± 1.1	3.90	6.870 ± 0.19	10.06	inf
2300	dsdv	56.5 ± 1.8	15.3 ± 1.4	225.1 ± 54.0	3.76	30.761 ± 1.04	22.42	inf
2500	anthocnet	89.0 ± 0.7	77.6 ± 3.9	974.6 ± 13.7	5.84	44.214 ± 1.24	121.09	inf
2500	aodv	77.3 ± 1.1	36.6 ± 2.4	786.2 ± 66.4	5.10	42.434 ± 1.01	49.97	inf
2500	olsr	67.0 ± 1.0	7.8 ± 0.6	30.6 ± 1.5	3.68	7.357 ± 0.20	9.97	inf
2500	dsdv	54.0 ± 1.5	16.0 ± 1.8	185.4 ± 42.4	3.58	31.533 ± 0.90	22.27	inf

Restated on CBR sources (#521)

Re-measured on constant-bit-rate sources. Until #521 every source ran ns-3's default 1 s on / 1 s off, so the pinned table above was taken at half the documented offered load. That table stays as the dated v1.3.0 record.

Provenance. main @ 3edbab6a, 3.42-opt, 900 s, range/disk PHY, 20 seeds per point, one point per scenario-matrix job: runs 37233283127 (1500), 37233284897 (1900), 37233286413 (2100), 37233288164 (2300) and 37233289720 (2500). Rescued as ../campaign/<run>-run.csv (aggregate) and <run>-run-runs.csv (per seed). scenario_check.py results: 0 FAIL; the WARNs are the #230 path-diversity window caveat and route-flap reorder notes. OLSR's PDR is offered-based (#510), so the estimate table at the top of this page does not apply to this block.

Attribution. This is not a CBR-only A/B. 3edbab6a also carries every change after v1.3.0 that the provenance note names (#327 pheromone weighting, #352 RNG streams), plus #510 (OLSR accounting) and

522 (TTL compensation). sweep_summary.py --vs the pinned CSVs reports

its baseline control as FAIL, as expected: every arm's offered load doubled. Quote this block's within-point orderings and paired deltas, not its difference from the pinned table.

Sweep area — mean of 20 run(s) per point, every baseline on identical realisations; ± is the 95% CI half-width (#293). Rendered with update-benchmarks.py's sweep builder from the rescued CSVs.

area long edge (m)	protocol	PDR % ±95	mean delay (ms) ±95	99th delay (ms) ±95	throughput (kbps)	NRL ±95	jitter (ms)	dOff90 (ms)
1500	anthocnet	97.2 ± 0.2	18.4 ± 0.7	247.0 ± 10.0	12.80	9.660 ± 0.26	27.64	47.2
1500	aodv	86.9 ± 0.6	24.4 ± 0.7	380.1 ± 13.1	11.45	27.531 ± 0.38	36.85	inf
1500	olsr	73.2 ± 1.1	6.1 ± 0.5	23.6 ± 0.8	9.64	2.620 ± 0.04	7.94	inf
1500	dsdv	69.6 ± 0.9	14.7 ± 0.8	357.7 ± 83.2	9.17	13.566 ± 0.17	23.25	inf
1900	anthocnet	95.0 ± 0.4	23.3 ± 1.0	321.4 ± 23.0	12.52	11.293 ± 0.34	34.93	137.3
1900	aodv	84.7 ± 0.7	22.0 ± 0.8	309.2 ± 19.0	11.16	24.679 ± 0.49	31.52	inf
1900	olsr	65.6 ± 1.3	6.1 ± 0.5	25.4 ± 1.0	8.64	3.026 ± 0.07	7.51	inf
1900	dsdv	62.3 ± 1.2	13.4 ± 0.7	258.3 ± 82.2	8.21	14.639 ± 0.30	20.24	inf
2100	anthocnet	93.6 ± 0.5	26.5 ± 1.5	381.9 ± 33.1	12.32	12.403 ± 0.47	39.50	166.2
2100	aodv	82.4 ± 1.0	22.8 ± 1.5	287.4 ± 27.6	10.85	23.749 ± 0.54	30.95	inf
2100	olsr	60.7 ± 1.6	6.5 ± 0.5	27.0 ± 1.3	7.99	3.313 ± 0.08	7.95	inf
2100	dsdv	58.3 ± 1.6	13.9 ± 1.0	221.7 ± 45.2	7.67	15.314 ± 0.41	20.23	inf
2300	anthocnet	91.7 ± 0.5	28.8 ± 1.0	454.9 ± 37.3	12.08	13.316 ± 0.35	42.75	inf
2300	aodv	81.0 ± 1.1	25.1 ± 1.8	378.6 ± 73.0	10.68	22.878 ± 0.50	32.53	inf
2300	olsr	57.3 ± 1.3	6.0 ± 0.5	27.6 ± 1.0	7.54	3.544 ± 0.08	7.06	inf
2300	dsdv	55.3 ± 1.2	14.0 ± 1.3	189.8 ± 41.2	7.28	15.810 ± 0.39	19.61	inf
2500	anthocnet	90.0 ± 0.8	30.2 ± 1.5	500.8 ± 47.1	11.85	14.088 ± 0.53	44.81	inf
2500	aodv	79.8 ± 1.3	28.1 ± 1.8	492.5 ± 87.0	10.51	21.933 ± 0.53	34.47	inf
2500	olsr	54.6 ± 1.8	6.1 ± 0.6	27.8 ± 1.7	7.19	3.756 ± 0.12	7.39	inf
2500	dsdv	52.7 ± 1.4	14.7 ± 1.6	226.9 ± 101.5	6.94	16.231 ± 0.43	19.64	inf

Paired per-seed deltas (mean ± 95 % CI half-width, 20 seeds; ns = the CI spans zero):

long edge (m)	ΔPDR anthocnet − aodv (pp)	Δdelay99 anthocnet − aodv (ms)	ΔNRL anthocnet − aodv	ΔPDR olsr − dsdv (pp)
1500	+10.3 ± 0.6	−133 ± 13	−17.87 ± 0.41	+3.52 ± 0.47
1900	+10.3 ± 0.6	+12 ± 31 (ns)	−13.39 ± 0.49	+3.25 ± 0.58
2100	+11.2 ± 0.9	+95 ± 39	−11.35 ± 0.48	+2.37 ± 0.60
2300	+10.7 ± 1.2	+76 ± 92 (ns)	−9.56 ± 0.45	+2.00 ± 0.53
2500	+10.2 ± 0.9	+8 ± 76 (ns)	−7.85 ± 0.34	+1.86 ± 0.52

What holds and what changed versus the pinned table: - The delivery ordering holds: AntHocNet > AODV > OLSR > DSDV at every point. AntHocNet's lead over AODV is +10.2 to +11.2 pp, against +10.3 to +12.1 pp pinned. It leads OLSR by +24.0 to +35.4 pp and DSDV by +27.5 to +37.3 pp. - OLSR over DSDV is significant but small (+1.9 to +3.5 pp). On the pinned page, the #510 estimate put it within ~2 pp at 2100 and 2500 m. - AntHocNet's tail is no longer the worst. Its delay99 is 247 to 501 ms, against 878 to 975 ms pinned. It is below AODV's at 1500 m, level at 1900/2300/2500 m, and above only at 2100 m. Against DSDV it is below at 1500 m (−111 ± 87 ms), level at 1900 m, and above from 2100 m. It is above OLSR's at every point. - AntHocNet's NRL is below AODV's at every point. Pinned, the two crossed at 2500 m. NRL is per delivered packet, so doubling the offered data lowers every arm's NRL.